Guillermo Kulemeyer
Physicist and PhD student in Computational Intelligence
I develop deep-learning models for scientific problems involving structure and uncertainty. My current research focuses on predictive and generative models for RNA structure.
I study how these models learn structural representations, generalize to less familiar sequences, and represent multiple plausible solutions.
Publications
2026
Homology-aware cross-validation strategies for generalization assessment in RNA structure prediction
Bugnon, Kulemeyer, Gerard, Di Persia, Stegmayer & Milone · under review, bioRxiv preprint
2026
The Less the Better: Improving RNA Secondary Structure Generalization with Structural-Aware Subsampling
Manuscript in preparation
2025
Enhancing Deep Learning Prediction of Base-Pairing Probability in RNA Sequences Using Structural Diversity
Poster · 1st Argentine Workshop on Bioinformatics and Computational Biology
2024
Dissociative Adsorption of Methane on Pt(110)-(1×2) Using Machine Learning
Poster · 10th Surfaces Physics & Chemistry Meeting, CNEA
Questions Guiding My Work
Learning Structural Representations
How can neural models capture the relationships between RNA sequence and secondary structure?
Generating Plausible Alternatives
When a sequence admits more than one structure, how can generative models represent that diversity?
Understanding Generalization
How does performance change as test sequences become less similar to the data used for training?
Teaching
Alongside my research, I teach Neural Networks I and II in the Artificial Intelligence Engineering program. See my academic activity.